TEDB. FLANAGAN AND B. S. RABINOVITCH
724
Vol. 60
num, copper, arsenic, antimony, tin and silver. The acid exhalations of these fumaroles are noteworthy. The concentration of acid in the steam is 1600 not great, but the total quantity is enormous. Zies / gP 1400 has estimated that in one year the fumaroles of the Valley of Ten Thousand Smokes give off one and a .Ei l2O0 8 lo00 quarter million tons of hydrochloric acid and two 800 hundred thousand tons of hydrofluoric acid. The ultimate destination of these acids is of interest. 600 The sea is an adequate sink for the hydrochloric 400 acid, but not for the hydrofluoric acid. Huge 200 quantities of fluorine are locked up in sedimentary phosphate deposits as the mineral apatite, and such 0.05 0.20 0.40 0.60 0.80 deposits may be the ultimate sink for the huge Solubility in weight per cent. quantities of hydrofluoric acid given off by volcanic Fig. 13.-The solubility of quartz in su erheated steam action. at high pressures, ex ressed as weight of SiOl is the But finally, as the ages go by, the process of condensate. After d r e y and Hesselgesser. magmatic differentiation is complete. From a important processes of differentiation, which were magma originally basic in composition, first the orespecially well brought out by Zies'6 in his studies thosilicate minei als separate, then the metasiliof the fumaroles of the Valley of Ten Thousand cates and plagioclases crystallize, leading to a Smokes. The Valley contains about 2 cubic miles magma from which either a nepheline rock, or of rhyolitic pumice, roughly of granitic composition, more commonly a granite, crystallizes. Later is a which was deposited after having been blown stage of pegmatite formation, where the influence through the floor of the Valley. Many fumaroles of the accumulated water has become dominant, are located in this area, through which have come and the highly concentrated aqueous solutions congreat quantities of steam, from which were depos- tain impot tant quantities of trace elements. But ited important quantities of lead, zinc, molybde- the process is a continuous one, called by chemists fractional crystallization, by geologists, differentia(15) E. G. Zies, N d . Ueoo. Soc., Contributed Tech. Papera, Katmai tion. Ser., 1, no. 4, 79 pp. (1929); Cham. Rars., 28, 47 (1938).
=)I
i
/ ,/"
8
8
EXCHANGE AND ISOMERIZATION OF trans-ETHYLENE-dz ON NICKEL. KINETIC STUDIES' BYTEDB. FLANAGAN AND B. S. RABINOVITCH Contribulim from the Department of Chemistry of the University of Washington, Seattle, Washington Received September I.%,I966
Activation energies and orders of the pressure dependence of the nickel-catalyzed isomerization and exchange reactions of trans-ethylene-& have been determined. The observed activation energies for isomerization are 13.5 and -6.7 kcal. from 35-105" and 170-195", respectively; the activation energy for exchange is C1.5 kcal. greater for the same temperature intervals. Both reactiona proceed by the same mechanism, and their relative rates as well as the small difference in activation energies between isomerization and exchange (and other results) are explained on the basis of an isotope effect in the decomposition of an intermediate ethyl radical. The order of the resaure dependence for both reactions is 0.3 a t 53" rising to 1.0 a t 170'. The C-H and C-D bond rupture isotope effectias been measured from -78 to 400'. Various mechanisms are discussed.
The mechanism of ethylene hydrogenation and exchange on transition elements is still controversiaL2 Important evidence against a dissociative exchange mechanisrna~~ was the work of Conn and Twigg6 in which no exchange was found between ethylene-do and ethylene-d4 on nickel wire. Furthermore, although Beeck's results4 strongly suggest dissociative adsorption takes place, alternative mechanisms for adsorption (and self-hydrogenation) (1) Abstracted in part from a thesis submitted b y Ted B. Flanagan to the Graduate School in partial fulfillment of the requirements for the degree of Doctor of Philosophy a t the University of Washington. Work supported b y the Office of Naval Research. (2) D. D. Eley, Disc.Faraday Soc., No. 8 , 99 (1950); G. C. Bond, Quart. Rev., 8 , 279 (1954). (3) A. Farkas, Trans. Faraday Soc., 36, 900 (1939). (4) 0.Beeck, Disc. Faraday Soc., No. 8, 118 (1980). (5) C. K. T. Conn and C. H. Twigg. Proc. Roy. SOC.(London), B i l l , 70 (1989).
have been given wherein ethylene does not liberate hydrogen to the catalyst.6 Recently, Douglas and Rabinovitch' have shown that exchange does take place between ethylene-do and ethylene-d4 on nickel-kieselguhr and on nickel wire; with trans-ethylene-dt, cis-trans isomerization was also found. Earlier, Baxendale and Warburst* found a small amount of isomerization of elaidic ester to oleic ester on a platinum black catalyst in the absence of hydrogen, although they attributed this to impurities on the surface. Koimmi9 ( 6 ) D. D. Eley, Disc.Faraday Soc., No. 8, 34 (1950); for further discussion see B. M. W. Trapnell, "Chemisorption," Academic Press, New York, N. Y.,1955, p. 185. (7) J. E. Douglas and B. 9. Rabinovitch, J . Am. Chsm. Soc., 74, 2480 (1952). (8) J. H. Baxendale and E. Warhurst. Trans. Faraday Soc., 8 6 , 1186 (1940). (9) M. Koizurni, J . Chem. Soc. Japan, 68, 1512, 1715 (1942).
,
June, 1956
EXCHANGE AND ISOMERIZATION
showed that ethylene4 and propylene exchanged over nickel powder catalyst a t 45". I n all cases where exchange or isomerization of olefins has been observed in the absence of hydrogen, the catalysts have had large surface areas as opposed t o the small areas involved where no reaction was found.6J0 Very recently, Jenkins and Rideal" have presented very strong evidence in favor of dissociative adsorption of ethylene on nickel. Because of the intimate relation between isomerization, exchange, and hydrogenation reactions of olefins and the fundamental importance of these reactions, a detailed investigation of the kinetics of the exchange and isomerization reactions of pure trans-ethylene-d2 on nickel has been undertaken and is reported here. Experimental Reactants.--trans-Ethylene-dl, cis-ethylene-dz and ethylene-d4 were prepared as described previ~usly.~The transethylene-dz was 99.3% pure, with ethylene-dl as the major impurity. cis-Ethylene-dz was 98.9% pure, with the remainder ethylene-&. Ethylene-dc was 97.6%, with 1% ethylene-ds and 1AT0 ethane-ds as the principal impurities. The ethylene-& was Phillips Petroleum Co. research grade, used without further purification. Deuterium (99.570), obtained on A.E.C. allocation from the Stuart Oxygen Co., was passed over hot platinized asbestos and through liquid nitrogen traps. Catalysts.-The catalyst for the major portion of the work was a 0.75 mm. "Ballast" Ni wire obtained from the Wilbur Driver Co., stated to be at least 99.7y0 pure and probably better than Y9.90/o. A helix of wire 2930 cm. long was wrapped around a sealed off Pyrex tube and fitted snugly inside of another Pyrex tube vessel. The geometric Rurface area was 730 0311.2. Thc net volume of the reactor was 33.5 rc. The catalyst was initially activated by oxidation with 10 cm. oxygen at 500' for 15 minutes and then reduction with 30 cm. hydrogen a t 330" for two hours, followed by evacuation for a t least two hours a t 330"; more reproducible results were obtained if the wire was subsequently reactivated merely by reduction and evacuation at 330'. The catalyst was protected at all times from stopcock grease and mercury by a COZtrap. A nickel film was prepared by evaporation of wire (greater than 99.9% ure) onto Pyrex glass a t 0'; it had an area of 3400 cm.2. $he film was protected by a COZt r y . A few experiments were performed with a Ni-kieselguhr catalyst; this was the same catalyst as was used earlier,' and was activated in the same manner, followed by evacuation a t 400' for three hours. Apparatus and Procedure.-A Pyrex glass vacuum apr t u s was used; the stopcocks were greased with A iezon The catalyst chamber was heated externally gy an electrical furnace and temperature was maintained constant to within one-half degree. Kinetic runs were made in the following way: a sample of trans-ethylene-dz was exposed to the catalyst for a given time, removed and analyzed for cis-ethylene-dz; a small fraction (approximately 1 %) was simultaneously removed for mass-spectral analysis; the remainder of the sample was frozen back into the reactor and reacted for an additional period. Usually six points were obtained in each run. The catalyst chamber was evacuated between runs. A few experiments on mixtures of trans-ethylene-& and excess ethylene-dr were carried out similarly. Diffusion to the catalyst was never a factor. The rate of reaction immediately after activation of the catalyst was very fast, as was also the rate of decrease of activity, and the catalyst was partially poisoned or seasoned, % . e . , exposed to ethylene-d? for a number of hours, usually a t 75', in order to increase reproducibility. Although the activity of the seasoned catalyst decreased further upon long exposure to ethylene, frequently the activity could be maintained .constant for a sequence of three, and on some occasions BIX, runs.
.
(IO) D. 0. Schissler, Ph.D. Thesis, Princeton University, 1951: T.I. Taylor and V. H. Dibeler, T H I SJOURNAL, 66, 1036 (1951). (11) G. I. Jenkins and E. Rideal, J . Chem. Soc., 2490, 2496 (1955): cf. Trapnell (ref. 0).
OF
fTUnS-ETHYLENE4 ON NICKEL
725
The runs on the determination of the temperature coefficient and the order of the pressure dependence of the rate were always made in series of three; the first and the third experiments were done under the same conditions and only those runs where the activity had not changed appreciably were accepted. This arduous procedure was adopted to eliminate any uncertainty that might attend the otherwise necessary correction of the data to standard catalyst activity, and in fact provided an experimental justification for such a correction. The runs on the temperature coefficient were performed a t constant concentration of ethylene in the gas phase. Analvsis .-The conversion to cis-ethylene-dz was followed by the transmission of the 842 cm.-l band of cisethylene-& using a Beckman IR-2 spectrophotometer. A calibration curve of per cent. transmission versus per cent. cis-ethylene-&, at a given total pressure, was constructed. As reaction proceeded the analysis became less accurate due to the presence of other isotopic olefins. The effect of the other constituents on the cis-analysis was estimated by adding ethyIene-d, and ethylene-& to cis-trans-ethylenedz in relevant amounts (ethylene-d1 has the largest effect) and obtaining corrected calibration curves. This correction proved negligible in the early stages of the reaction. A calibration curve of cis and trans-ethylene-dz mixtures containing various percentages of added ethylene-dc was constructed and used for a few special runs. Other ethylenic species were analyzed at 70 volts with a model 21-103 Consolidated Mass Spectrometer. The patterns of ethylene-do, ethylene-dl, cis-ethylene-&, transethylene-dz, asym-ethylene-dz and eth lene d, were determined using ethylenes available in our raboitory, and that of ethylene-d8 from the literature.l2 Analysis for a s p ethylene-dz was made by use of the 16 peak, because the parent peak was interfered with by Cis- and trans-ethylenedt. Arialysis for asym-ethylene was not as accurate as that for the other ethylenes.
Results and Discussion Possible Effect of Residual Hydrogen Left on the Catalyst from the Activation Procedure.Experiments were done t o assess the influence of residual hydrogen. After reduction with deuterium and evacuation at 330°, an amount of ethylene-do corresponding to one monolayer was added to the catalyst a t 100" for a period of time longer than the usual reaction time; negligible ethylene-dr was observed while a larger percentage of self-hydrogenation, ethane-&, was found, analogous to the earlier result of Douglas and Rabino~itch.~ The exchange of trans-ethylene-dz was also studied on a fresh Ni film on which no H can be present. Complete equilibration was observed in less than one hour a t 23", with small amounts of various ethanes also occurring. A lower rate of isomerization and exchange was found at -78". Determination of Rate Constants and Time Reaction Order.-Rate constants for isomerization and exchange reactions were calculated from first-order expressions. The equation for isomerization was dCc/dt = ICi(C2" - 2C3 - 2Cc), where C, refers to an ethylene with n deuterium atoms, C, refers to &-ethylene-& and Czo refers to the initial trans-ethylene-d2. This approximate equation is appropriate for small amounts of reaction.13 For most runs, reaction was allowed to proceed to give finally only around S-lO% total of ethylene-& plus ethylene-& (single step exchange products) , and about 8-12% cis-ethylened2, de(12) V. Dibeler, F. Mohler and M. DeHernptinne, Bull. Nal. Bur. 63, 107 (1954). (13) T. B. Flanagan, Thesis, University of Washington, 1955; for the sake of brevity in this paper, reference will sometimes be made to this thesis where the indicated information may be found. S(dS.,
TEDB. FLANAGAN AND B. S. RABINOVITCH
726
pending upon the temperature; higher exchange products were negligible (indicating that exchange went by a stepwise process). A correction was made for the major source of error in the equation, i.e., cis-ethylene-dz reaction to give exchange products. The rate constants for exchange, k e l were calculated from the simple approximate equation, d(C1 C3)/dt = k e ( C z o - (C, C,)),which holds for the small percentage reaction employed. Experimentally, ethylene41 and ethylene-d3 are essentially equal at all times. Frequently, as the reaction proceeded, the firstorder constants for both exchange and isomerization fell somewhat, in a parallel manner. Following a Tun during which the values of k fell, after evacuation the succeeding run sometimes started off at the original level of catalyst activity. Because of the fall-off present in some runs, average k values were not always satisfactory. Instead, a curve was fitted to the earlier portion of a run and a representative value of k was calculated for a given small percentage reaction where the fall-off, if any, was small. A typical run is shown in Table I. Isomerization was faster than total exchange at all temperatures (-80-400°).
+
+
Vol. 60
constant over each pressure range investigated and could be averaged over all groups of runs at a particular temperature since the order appeared to be independent of variation of catalyst activity (twofold) from one group to another; obviously, correction of the data to a standard catalyst activity would give identical results. ORDEI~ OF Temp.
("C.)
THE
TABLE I1 PRESSURE DEPENDENCE OF ISOMERIZATION AND EXCHANGE REACTION Isomerization
Exchange
"n"
"n"
Pressure variation (mm.)
53 0.3" 0.3" 17-68 104 0.4" 0.4" 20-160 129 0.68 0.57 2 1-86 152 0.75 0.84 22-90 172 1.0 1.0 24-94 a Irreproducibility and experimental error were greatest at low temperatures and the cited value, an average of a limited number of measurements, may contain considerable error.
Within experimental error the pressure dependence order for isomerization is equal to that for exchange at a given temperature; both rise with increase of temperature. Activation Energies.-Representative IC values TABLE I for isomerization and exchange were evaluated a t TYPICAL RUN, #66B (75"C., 7.4 CM.) 7.5% isomerization and 3y0 ethylene-&; if k values calculated at other percentages of reaction than these were used, essentially the same activation 1.95 5.5 4.8 3.1 2.69 13 2.0 energies were obtained.13 The determination of 2.9 2.43 21 3.1 3.0 7.8 4.4 the activation energies was made in small groups 9.5 4.5 2.9 2.53 3.6 26 3.4 of runs (see Procedure) where each group was at 2.49 4.2 10.7 4.4 2.8 31 4.1 constant catalyst activity. To represent all the 11.8 4.2 2.7 2.59 4.3 4.6 36 data on one Arrhenius plot, the data were corrected 2.62 4 . 2 2 . 6 5.2 13.1 42 5.6 to standard cat,alyst activity. The standard acIn spite of the fall-off in the first-order rate con- tivity toward isomerization was chosen as 7.5y0 stants observed during some runs, the isomeriza- isomerization in 22 minutes for 7.4 cm. of transtion and exchange reactions of trans-ethylene-& in ethylene-dz at 75", and that toward exchange was the absence of hydrogen are believed to be first 3% ethylene-d3 for the same time under the same order. The decline of the rate constants in some conditions. The observed activation energy for isomerizainstances is quite probably due to some self-poisoning process, because ethylene is known to poison Ni tion is 13.5 kcal. from 35 to 105"; it then drops ~ a t a l y s t s . ~ J Also, ~ ~ ~ *we have found good agree- continuously, reaching -6.7 kcal. for the interval ment between theoretical equations, based on 170 to 195". The observed activation energy for first-order elementary reactions, and experimental exchange is 15.0 kcal. from 35 to 105; it also decreases reaching -6.7 kcal. for the interval 170 to results for the equilibration of ethylenes.l5 Pressure Reaction Order.-The order of the 195" (Fig. 1). The average difference, AE = E, pressure dependence, n, was determined from the E i , is 1.5 kcal. for 35-105", and is also small, but slope of the plot of log rate us. log pressure at con- less accurately known, above 105". These results s tant catalyst activity. The pressure dependence refer to k values corrected to standard activity at a was studied from 53 to 170" within certain pressure constant concentration of 2.1 X lo1*molec. cc.-'; ranges. At reaction pressures lower than 3.7 E values calculated from each group of runs, for cm. (75') it became difficult to analyze for cis- which the catalyst activity was constant within a ethylene-dz and only exchange products were group but differed between groups, were concordanalyzed. The results in Table I1 are each an aver- ant.13 Correction of the data to a constant cataage of a number of groups of runs, each group being lyst activity thus appears to be a valid procedure at constant catalyst activity. The order appeared here also. The ratio of cis-ethylene-dz to ethylene-ds, at a (14) J. R. Anderson and C. Kernball, Proc. Roy. SOC. (London), given percentage reaction, depends only upon the 228A, 361 (1954). Beeck (ref. 4) has found C I to Ce and higher temperature and not on catalyst activity (Table polymers upon hydrogenation of a preadsorbed ethylene layer on a Ni film: we have foud ethane and a very small amount of butene upon 111). In the extreme, results on nickel films, Niaddition of what corresponds to a monolayer of ethylene to the Ni kieselguhr and a Pd catalyst bear this out, Le., the wire. ratio of isomerization to exchange at a given per(15) T. B. Flanagan and B. 6 . Rabinovitch, THIS JOURNAL, 60, 730 (19513). centage reaction was independent of the catalyst.
.
t I
-
tI
,
EXCHANGE AND ISOMERIZATION OF ~TWLS-ETHYLENE-~ ON NICKEL
June, 195G
C
rhenius plot of the ratio of early rates of isomerization and exchange (Fig. 2). AE by this method for the region below 105" is 1.1 kcal., in good agreement with the difference between E, and Ei calculated from the standard activity plot (Fig. 1).
I
0
0
.IO
0
: - I.(
727
0.7
4 €
0.6
$
-20
5 0.5 $!
4 Q
2 -3.0
& u
2
04
.0.3 20 Fig. 2.-AE
25
4 x io3
3.0
from early rate ratio of isomerization and exchange.
- 3s
I n the temperature range 35-105", the temperature independent factor, A , for runs on the seasoned catalyst brought to standard activity is 1.1 X 102 sec.-l and 16.0 see.-' for total exchange and isomerization, respectively. Mechanism of the Isomerization and Exchange -4.0 Reactions.-Since both isomerization and exchange reactions have the same kinetics, i.e., time order, pressure order, similar magnitudes of activation energies (the small difference is discussed later), and variation of these quantities with temperature, -41 no special mechanism can apply for isomerization. 2.5 3.0 3 It is concluded that both reactions occur by the bT x io3. Fig. I .-Activation energies for the isomerization and ex- same mechanism. It is desirable a t this point in the presentation of the results to show that the change of trans-ethylene-dn. reactions may be assumed to proceed via an Since the ratio of isomerization to exchange was adsorbed ethyl radical. independent of catalyst activity, AE could be calcuBy reason of the experimentally observed ratio lated directly from all groups of runs by the Ar- of isomerization to exchange the following mechanisms cannot, of themselves, account for the resulbs: TABLEI11 the Farkas mechanism3 RATIO OF ISOMERIZATION TO EXCHANGE AT VARIOUS TEMPERATURES A N D PRESSURES. SOMETYPICAL RESULTS CZH, HzC=CH + 5 Temp. ("C.)
Relative cat.
activ.
EdEi
Press. (mm.)
(at 2%
Ea)
54 54 54 54 54 54
11.1 2.2 0.31 0.27 0.27 0.27
6.9 6.9 6.9 3.4 6.9 13.8
3.30 3.15 3.42 3.40 3.50 3.40
75 75 75
1.1 0.81 0.44
7.4 7.4 7.4
2.80 2.69 2.75
the Bee& mechanism4 CyHd
H$=C*H
+25
a radical dissociative mechanism G H ~_r HC-CH~
+5
* * involving a radical postulated as an intermediate in acetylenic hydrogenationI6; also a disproportions,(16) J. Sheridan, J . Chem. Soc., 133 (1945L
TEDB. FLANAGAN AND B. S. RABINOVITCH tior1 reaction among associatively adsorbed species like, 2HzC-CH2 F? C2Hb HC=CH2. * * On the other hand several mechanisms which proceed with opening of the double bond via an adsorbed ethyl radical are consistent with the results. These are the "associative" mechanism(A)1 7 3 proceeding through an adsorbed ethyl radical or half-hydrogenated state (h.h.s.)
+
A,
and a Rideal mechanism (R) R,
for which there has recently been proposed further strong support.llmls I n the absence of gas phase hydrogen or deuterium, the H or D pool necessary for these mechanisms must arise from dissociative adsorption. A small amount of H or D formed by dissociative ad* * sorption can catalyze the entire reaction sequence; removal of H by any process signifies poisoning of the surface and decline in rate. C-H/C-D Bond Rupture Probabilities in the Half-hydrogenated States.-The ratio of the C-H to C-D bond rupture probabilities, cy, for each h.h.s. which arises by addition of an H or D atom to trans-ethylene-&, HDC-CH2D (I) and HDC* * CHDz (TI),is determinable by use of the equations for Cc and C3concentrations as a function of time, developed in the following paper,16namely C,
-5Oe-Na -I-11,
Vol. 60
and
where it is assumed that the transmission coeficients and symmetry numbers cancel, that the force constant of the breaking bond is zero, that the rest of the bonds in the activated complexes are identical, except for an extra C-H bond when a G D bond breaks and an extra C-D bond when the C-H bond breaks and where only the bond stretching frequencies have been retained. This equation also applies for the isotope effect in the reactions of halfhydrogenated state 11, and hence a! has the same value for I and 11. At low temperatures, E, should equal AEO, the zero point energy difference for C-H and C-D, 1.14 kcal. Calculated values from the equation are high compared to the experimental values which approach the limiting value 1.36 at 400'. Complication of the model to obtain concordance with the data is not worth while here. Gas phase studies of ethyl radical decomposition are in progress. Resolution of Total Isomerization into Component Reactions.-The total isomerization at any temperature may be resolved into the contributions from half-hydrogenated states I and 11. I n the eaYly reaction, ethylene-dl arises from I and ethylene-& arises from 11. From I, the rates of ethylene-dl and cis-ethylene-d2 production are
+ 1) + 12.5 + 25e-81 + 12.5e-Wt
where kt is the total specific rate of decomposition - 25e-281 of I and (C-H)1 and (C-H)Z are the fractional probfrom the product analysis at any time. As shown abilities of C-H rupture in I and 11, respectively; (C-D), and (C-D)2 are the corresponding quantities below, a! is the same for I and 11. Values of a in a given run are independent of per- for C-D rupture, and the l/2 factor takes account of centage reaction (Table I). The experimental val- re-formation of trans-ethylene-dz. From I1 ues in Table IV are the average of several runs; - = kt'(C-H)2(II) = kt'(a/(2 + a))(II) =
and C, = 25
for each run, three or more determinations of cy were dt made a t different times. An Arrhenius plot for a! k' kt+ '(W = (GD)a(II) = 2 a (2) shows straight line behavior below 160" and the slope gives E, = 1.5 kcal. (Fig. 3). where kt' is the specific rate of decomposition of 11. A frequently used, simple model of the isotope efFrom these equations and the relation between fect for the intramolecular decomposition of h.h.s. I kt and kt', namely k( = (a! 2)/(2a! l)kt, it is and I1 may be formulated using equa.tions of Bigel- readily shown that the ratio of the concentrations of eisen.20 For h.h.s. I, the reactions are I and I1 is a! at any temperature; also the isomeriH k H H zation produced from I relative to that from 11 is DCH +DC-CH + D or C Z H ~ D + TI* * * * a2. Table IV summarizes the calculations. Isomerization and Exchange Experiments with k' H H H D +DC-CD + 5 or C ~ H ~ D +Z7 Excess Ethylene-d4.-At small times the normal * * * ratio of C,/(Cl C3) at any temperature is, from (17) J. Horiuti and M. Polenyi, Trans. Faradav Soc.. 80, 1164 (1934). 1)/2a!-(3); if only I equations 1 and 2, (a2 (18) A related meahanism has been proposed by Markham, Wall or I1 were present the ratio would be a! or l/a, and Laidler, ((a) J . Chem. Phys.. 80, 1331 (1952): (b) K . J. Laidler, respectively. If the ratio of 1/11 were changed "Catalysis," Vol. I. Ed. P. Emmett, Reinhold Publ. Corp.. New relative to the normal value, the ratio of isomerimYork. N. Y.. 1954), H~C-CHI + C2Hs # CzHs + HzC-CHr. and a * * * * tion to exchange would shift in the direction indivariant, intermediate between the Laidler and associative mechanisms. cated. Gaseous Hz or Dz was used to alter the has been suggested by Wilson, Otvos, Stevenson and Wagner ((c) Ind. Eng. Chem., 4S, 1480 (1953)); for present purposes these may be normal ratio of 1/11. Such experiments gave included in mechanism A. qualitative agreement with the expected shifts in (19) A modified Rideal mechanism, C ~ H I+ HIC-CHI # CiH, + * * * products, but the experiments were complicated HC-CHI, is ruled out by the same experimental work of Jenkins and by hydrogenation. * * Rideall1 which supports a RideaI type mechaniern. A source of excess H or D atoms was obtained by
%
%
+
1
(20) J. Bigeleisen, J . Chem. Phys., 17, 675 (1949).
+
+
*
*
+
EXCHANGE AND ISOMERIZATION OF tTUnS-ETHYLENEd ON NICKEL
June, 1956
addition of ethylenedo or ethylenedd. This method was free from the complication of hydrogenation. For swamping concentrations of ethylene-dr or ethylenedo with truns-ethylene-dt, only I1 or I arises. The effect on the products is more dramatic for the case of only I1 present; e.g., a t 90" where LY = 2.4, 1/a = CC/Ca should be 0.42, compared to the normal ratio 1.42. In two experiments a t 90°, 3.8 cm. of trans-ethylene-dz and 12.9 cm. of ethylened4 gave (after correction for ethylenedl which arose from ethylened4) an average ratio, CC/(C, Ca), of 0.60, Le., approaching the limiting value 0.42. The expected ratio, CJ(C1 C3), may be calculated: for the ethylened4-tran,s-ethylened2 ratio of 3.4 :1 used, assuming the ethylenes are adsorbed in the ratio of their pressures, (I)/(II) is 0.3. The general expression for the ratio, isomerisation/exchange, isla (a(I/II) l)/(I/II) a),which reduces to equation 3 for 1/11 = LY (pure truns-ethylenedz); for 1/11 = 0.3, isomerization/exchange = 0.67 which agrees with the experimental value. Composition of Observed Activation Energies for Isomerization.-The fact that the ratio 1/11 varies with temperature should be considered. Table IV shows that 92.4% isomerization takes place via I a t 35". At this temperature, therefore, TABLE IV TEMPERATURE DEPENDENCE OF C-H/C-D RUPTUR~
729
+
+
+
RATIO,LY,
+
AND
RELATED QUANTITIES
a
Fraction of I = I/(I 11)
'
0.1 I .o
1
I
2.o
3.0
4.0
5.0
'/T lo3 Fig. 3.-Temperature dependence of a.
Conclusions.-For reasonabIe values of a number of parameters, both the Rideal and associative mechanisms appear to fit the experimental results 0.996 15.9 0.94 -78 (pressure order and its variation with temperature, .98 .87 6.5 -25.5 decrease of activation energy with temperature, .96 .83 4.9 0.0 stepwise nature of the exchange, and the rate of .92 3.3 .77 34.6 reaction), as shown by absolute rate theory cal.89 2.9 -74 43.0 culations with these mechanisms.21 The details .89 2 . 8 .74 53.0 are not presented; we are unable to distinguish the .88 .73 74.0 2.7 validity of these mechanisms, and the possibility .87 2.6 .72 84.0 exists that some or all of the observed temperature .85 .71 2.4 104.0 of parameters may have its origin in a type variation .79 1.95 .66 148.0 of experimental artifact, such as changing H atom .77 1.80 168.0 .64 concentration on the surface due to shift with tem.76 1.77 195.0 .64 perature of the "poisoning" r e a c t i ~ n s . ~ ~ItJ ~may .73 1.64 .62 315.0 be noted that agreement of calculated rates with .66 1.40 .58 429.0 experiment (calculated for 27") requires values of e', AE = E, - E i should be almost equal to E, (1.5 H coverage, and e", h.h.s. coverage of there kcal.). I n general, the value of AE a t any temper- is evidence that these are reasonable magnitudes. la ature islS Acknowledgments.-We thank Dr. J. H. SingleLY2 - 1 Ea ton for his kind assistance in the nickel film experiA E = aP/(1 + LY')E,- 1/(1 + d ) E a = CY2 + 1 ments and for the use of his apparatus; also from which AE = 1.2 kcal. a t 35", dropping to 0.8 Professor A. L. Crittenden and Mr. B. J. Nist for kcal. a t 170". The experimental accuracy was not some of the mass-spectral analyses and the ONR sufficient to reproduce the fall in AE (see earlier for their support. Results). (21) Some of the calculationa are given in ref. 13. Temp. ("C.)
+
Fraction of isomerization from I
